Synthesis of window functions for reducing systematic errors of multiplexed fiber-optic sensors


In the paper, the systematic (bias) error of digital processing of the biharmonic signal of two multiplexed fiber-optic interferometers under spectral interrogation, when the registered spectral transfer function being processed through a discrete Fourier transform has been analyzed. The synthesis of a special weighted window function was put forward; this made it possible to reduce the systematic errors in determining the frequencies of a polyharmonic signal without a significant increase in the random (noise) error. The effectiveness of this approach was proven through numerical simulation and experimentally by comparison with the error when using standard Dolph – Chebyshev window. The proposed approach can be applied in any problems related to the estimation of frequencies and phases of polyharmonic signals.